Semiconductor manufacturing system daily output prediction based on phase space reconstruction

In order to manage and control semiconductor wafer fabrication system (SWFS) more effectively, the daily output prediction data of wafer fabrication are often used in the planning and scheduling of SWFS. Because of nonlinear certainty and stochastic character of the daily output time series, an artificial neural network prediction method based on phase space reconstruction and ant colony optimization is proposed, in which the chaos phase space reconstruction theory is used to reconstruct the daily output time serials, the neural network is used to construct the daily output prediction model, the ant algorithm is used to train the weight and bias values of the neural network prediction model. Through testing with factory production data and comparing with traditional prediction methods, the effectiveness of the the proposed prediction method is proved. ©2009 Journal of Mechanical Engineering.